基于模糊聚类方法的盲文本水印算法研究

Blind Text Digital Watermarking Research Based on Fuzzy Clustering

利用模糊聚类理论及汉字的属性特征,提出了一种针对中文文本文档的盲水印算法。根据人类视觉系统的特点,将汉字按照汉字结构类型、汉字笔画数和使用频率进行分类,选择其中汉字笔画数较少、使用频率较高的类别进行水印嵌入,水印检测无须原始文档。实验表明该算法在面对文本的局部删减、添加、格式调整攻击时具有良好的鲁棒性。

This paper proposes a blind text watermarking algorithm by utilizing fuzzy clustering theory and Chinese character＇ s properties. According to HVS, it classifies the Chinese characters based on their types of structures, strokes and frequency. Then is chooses the Chinese characters of less strokes and more frequencies to be embedded. In detecting the watermarks, original text is not necessarily needed. Experimental results show that the algorithm can achieve the robustness, which is resistant to the attacks of partial deleting, adding, and adjusting the format of text.